Reply to Dummett’s Comment

Abstract

The problem with identifying truth with verification is that, apart from mathematical sentences, verification (or warranted assertibility) is “a property that may be lost.” Michael concedes this is so in the case of sentences about photons, on account of what he refers to as “the instability of scientific theory” but it seems to me to be equally so in the case of sentences about tables and chairs. If “there is a table in the next room” is held to be verified in experiential circumstances C, then, no matter how circumstances C may be spelled out by the Verificationist (assuming we are given some finite description), we can easily tell a story consistent with physical theory about how we might be caused to have those very experiences even though no table was present in the next room. Worse still, for the Verificationist, once such a story is told, we can easily think of ways that we could verify it if it were true, though not necessarily at the same time that the experiential circumstances C obtain. So, on his own grounds, the Verificationist must regard it as possible that this story should on some occasion be true, and hence regard it as possible that the circumstances C obtained although there was no chair in the next room.